Meta on Tuesday introduced Muse, a personal AI agent aimed at everyday consumer tasks in the U.S., TechCrunch reported. The product is separate from Meta’s earlier Muse Spark model releases: Muse is the agent surface that connects to a user’s apps and tries to do work, not just answer questions.
According to TechCrunch, Muse can send email, book travel, lower bills, fill forms, turn recipe Reels into grocery lists, send invitations, and make purchases, with checkout through Link by Stripe at launch and Shop Pay and 1Password integrations said to be coming. Users opt in to each app or service one at a time. When a service has a public API, Muse can connect with credentials the user provides; when it does not, Muse can use a browser instead. The agent is powered by Meta’s Muse Spark model.
Availability at launch is the web at muse.ai, iOS and Android apps, and WhatsApp chats, with Meta AI glasses listed as a later surface. Muse is free to start but requires a payment card. Paid plans are Power at $20 a month and Maximum at $100 a month for higher usage; Meta expects most people to stay on the free tier and says the app shows a usage meter before charges kick in. The agent can keep working after the user leaves the app and can make unprompted suggestions based on prior conversations, TechCrunch said.
Meta’s trust pitch centers on Muse Secure VM, described as a dedicated virtual machine with its own browser, plus a separate Sentinel agent kept apart at the system level. Meta claims Muse does not see passwords or payment methods and does not share conversations or data with Meta’s ads systems. TechCrunch noted those claims need outside scrutiny and set the launch against Meta’s privacy history, including FTC actions and the recent multistate social-media settlement.
This is a consumer agent launch day, not another model-card drop. Muse Spark 1.3 already shipped for developers; Muse is Meta asking people to hand an agent the keys to mail, calendars, and wallets.